Section 1

Various Propositions

[14/100%]
Show
LaTeX
Problem 4.164

If the function f(z)f(z) is regular, single valued and bounded in the strip 0≤ℜz≤π0 \leq \Re z \leq \pi, and vanishes there at the points z1,z2,z3,…,zn,…z_{1}, z_{2}, z_{3}, \ldots , z_{n}, \ldots (zn=xn+iynz_{n}=x_{n}+iy_{n}) then either the series of positive terms

e−∣y1∣sin⁡x1+e−∣y2∣sin⁡x2+e−∣y3∣sin⁡x3+⋯+e−∣yn∣sin⁡xn+⋯ e^{-\left|y_{1}\right|} \sin x_{1} + e^{-\left|y_{2}\right|} \sin x_{2} + e^{-\left|y_{3}\right|} \sin x_{3} + \cdots + e^{-\left|y_{n}\right|} \sin x_{n} + \cdots

is convergent, or f(z)f(z) is identically equal to 00.

?
Remark.
?

See Problem 297 of Part III.

Problem 4.165

Assume that the coefficient f1(z)f_{1}(z) of the linear homogeneous differential equation

y(n)+f1(z)y(n−1)+f2(z)y(n−2)+⋯+fn(z)y=0 y^{(n)} + f_{1}(z) y^{(n-1)} + f_{2}(z) y^{(n-2)} + \cdots + f_{n}(z) y = 0

is an entire function. The necessary and sufficient condition that the general solution of this differential equation be an entire function is that the remaining coefficients f2(z),f3(z),…,fn(z)f_{2}(z), f_{3}(z), \ldots , f_{n}(z) are also entire functions.

?
Problem 4.166

Let the function w=φ(z)w=\varphi (z) be regular (possibly many-valued) in the annular region 0<∣z∣<ρ0 < \left|z\right| < \rho, ρ>0\rho > 0, and satisfy there identically an equation of the form

F0(z)wl+F1(z)wl−1+⋯+Fl−1(z)w+Fl(z)=0, F_{0}(z) w^{l} + F_{1}(z) w^{l-1} + \cdots + F_{l-1}(z) w + F_{l}(z) = 0,

where F0(z),F1(z),…,Fl−1(z),Fl(z)F_{0}(z), F_{1}(z), \ldots , F_{l-1}(z), F_{l}(z) are regular in a neighbourhood of the point z=0z=0. If there exists a power series c0+c1z+c2z2+⋯c_{0}+c_{1}z+c_{2}z^{2}+\cdots such that the function (φ(z)−c0−c1z−c2z2−⋯−cn−1zn−1)z−n(\varphi (z)-c_{0}-c_{1}z-c_{2}z^{2}-\cdots -c_{n-1}z^{n-1})z^{-n} remains bounded in the neighbourhood of z=0z=0 for infinitely many values of nn, then φ(z)\varphi (z) is regular in the neighbourhood of z=0z=0, and we have φ(z)=c0+c1z+c2z2+⋯+cnzn+⋯\varphi (z)=c_{0}+c_{1}z+c_{2}z^{2}+\cdots +c_{n}z^{n}+\cdots.

?
Problem 4.167

Let the function f(z)f(z) be regular and single-valued for R≤∣z∣<∞R \leq \left|z\right| < \infty, R>0R > 0. Then there exists an integer p≥0p \geq 0, an entire function G(z)G(z) and a power series

ψ(z)=c−1z+c−2z2+c−3z3+⋯ , \psi (z) = \frac{c_{-1}}{z} + \frac{c_{-2}}{z^{2}} + \frac{c_{-3}}{z^{3}} + \cdots ,

convergent for ∣z∣≥R\left|z\right| \geq R, such that for R≤∣z∣<∞R \leq \left|z\right| < \infty we have

f(z)=z−pG(z)eψ(z). f(z) = z^{-p} G(z) e^{\psi (z)}.

If in the assumption as well as in the conclusion we replaced the inequality R≤∣z∣<∞R \leq \left|z\right| < \infty by the inequality R<∣z∣<∞R < \left|z\right| < \infty, the theorem obtained would be invalid.

?
Problem 4.168

Let f(z)=f(x+iy)f(z)=f(x+iy) be a meromorphic periodic function of period 2π2\pi that has in the strip 0≤ℜz<2π0 \leq \Re z < 2\pi only a finite number of zeros and poles. Denote by M(y)M(y) the maximum and by μ(y)\mu (y) the minimum of ∣f(x+iy)∣\left|f(x+iy)\right| for 0≤x≤2π0 \leq x \leq 2\pi. If it is the case that

lim sup⁡∣y∣→∞log⁡log⁡M(y)∣y∣<1or thatlim inf⁡∣y∣→∞log⁡log⁡μ(y)∣y∣>−1, \limsup _{\left|y\right|\to \infty } \frac{\log \log M(y)}{\left|y\right|} < 1 \qquad \text{or that} \qquad \liminf _{\left|y\right|\to \infty } \frac{\log \log \mu (y)}{\left|y\right|} > -1,

then f(z)f(z) must be a rational function of eize^{iz}.

?
Problem 4.169

Let f(z)f(z) be an analytic function, z=x+iyz=x+iy, and let the square of the absolute value of f(z)f(z)

∣f(z)∣2=φ(x,y) \left|f(z)\right|^{2} = \varphi (x,y)

be an algebraic function of the real variables xx and yy. Then the function f(z)f(z) is itself an algebraic function of zz.

?
Remark.
?

H. A. Schwarz, oral communication.

Problem 4.170

There does not exist a function which is regular analytic along the real axis and which assumes for real values of the variable every value in the interior of a fixed circle. In brief: There does not exist an analytic Peano-curve.

?
Remark.
?

E. Landau.

Problem 4.171

Find, if possible, n+1n+1 analytic functions f1(z),f2(z),…,fn(z),f(z)f_{1}(z), f_{2}(z), \ldots , f_{n}(z), f(z), which differ from each other not merely by constant factors, are regular in a domain D\mathfrak {D} and for which the relation

∣f1(z)∣+∣f2(z)∣+⋯+∣fn(z)∣=∣f(z)∣ \left|f_{1}(z)\right| + \left|f_{2}(z)\right| + \cdots + \left|f_{n}(z)\right| = \left|f(z)\right|

holds in D\mathfrak {D}.

?
Problem 4.172

Find, if possible, two non-constant analytic functions f(z)f(z) and g(z)g(z), which are regular in a domain D\mathfrak {D} and for which the relation

∣g(z)∣=ℜf(z) \left|g(z)\right| = \Re f(z)

holds in D\mathfrak {D}.

?
Remark.
?

See Problem 58 of Part III.

Problem 4.172.1

The modular graph (see Volume I, p. 130) of an entire function g(z)g(z) is everywhere convex from below in just two cases: Either when

g(z)=(az+b)n g(z) = (az+b)^{n}

or when

g(z)=eaz+b, g(z) = e^{az+b},

where aa and bb are arbitrary complex numbers, nn a non-negative integer.

?
Remark.
?

See the definition preceding Problem 130 of Part III. Cf. G. Polya, problem: Jber. deutsch. Math. Ver., Section 2, 35, p. 48 (1926). Solution by H. Kneser, op. cit. 35, p. 119 (1926), and by N. Tchebotarev, op. cit. 36, pp. 40-41 (1927).

Problem 4.173

We say that two entire functions f(z)f(z) and g(z)g(z) have the same aa-points if the two functions

f(z)−ag(z)−aandg(z)−af(z)−a \frac{f(z)-a}{g(z)-a} \qquad \text{and} \qquad \frac{g(z)-a}{f(z)-a}

are both entire. Find two distinct entire functions, which have the same aa-, bb- and cc-points. (Naturally we require b≠cb\neq c, c≠ac\neq a, a≠ba\neq b.)

?
Remark.
?

Note that there do not exist two distinct entire functions of finite order that have the same aa-, bb-, cc- and dd-points (a,b,c,da,b,c,d all distinct). G. Polya: Nyt Tidsskr. for Math. (B) 32, p. 21 (1921).

Problem 4.174

Does there exist an entire function G(z)G(z) which satisfies the equations

G(0)=a0,G′(1)=a1,G′′(2)=a2,…,G(n)(n)=an,… G(0)=a_{0}, \qquad G'(1)=a_{1}, \qquad G''(2)=a_{2}, \qquad \ldots , \qquad G^{(n)}(n)=a_{n}, \qquad \ldots

if the sequence of numbers a0,a1,a2,…a_{0}, a_{1}, a_{2}, \ldots is prescribed arbitrarily?

?
Remark.
?

For the analogous interpolation problem for polynomials cf. Problems 75, 76 of Part VI.

Problem 4.174.1

Given the numbers

zn;mn;wn,0,wn,1,…,wn,mn−1 z_{n}; \quad m_{n}; \quad w_{n,0}, w_{n,1}, \ldots , w_{n,m_{n}-1}

for n=1,2,3,…n=1,2,3,\ldots subject to the conditions:

zk≠zlifk≠l, z_{k} \neq z_{l} \qquad \text{if} \qquad k\neq l, lim⁡n→∞∣zn∣=∞, \lim _{n\to \infty } \left|z_{n}\right| = \infty ,

mnm_{n} is a positive integer, however the wn,kw_{n,k} are unrestricted.

Then there exists an entire function such that

f(k)(zn)=k! wn,k f^{(k)}(z_{n}) = k! \, w_{n,k}

for n=0,1,2,…n=0,1,2,\ldots, 0≤k<mn0\leq k<m_{n}.

?
Remark.
?

The condition prescribes the mnm_{n} initial terms of the expansion of f(z)f(z) about the point znz_{n}

f(z)=wn,0+wn,1(z−zn)+⋯+wn,mn−1(z−zn)mn−1+⋯ f(z) = w_{n,0}+w_{n,1}(z-z_{n})+\cdots +w_{n,m_{n}-1}(z-z_{n})^{m_{n}-1}+\cdots

for n=1,2,3,…n=1,2,3,\ldots.

Problem 4.174.2

Suppose that F(z)F(z) and G(z)G(z) are entire functions that have no common zeros. Then there exist two entire functions f(z)f(z) and g(z)g(z) such that

F(z)f(z)+G(z)g(z)=1. F(z)f(z)+G(z)g(z)=1.
?